JP3283208B2 - Automatic focus control method and automatic focus control system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION The present invention relates to an automatic focus control of an optical microscope or an electron microscope, and is used, for example, in a length measuring system (dimension SEM) for measuring a dimension of a fine pattern.

[0002]

2. Description of the Related Art In a semiconductor manufacturing line, a dimensional SEM device using an electron beam has been introduced in order to measure and manage the size of a fine pattern. This type of dimension SEM
Currently, the apparatus is fully automated, and first performs an automatic focus adjustment process for each measurement point on a wafer, and then measures a pattern dimension. The reason why the automatic focus adjustment processing is performed for each measurement point is mainly that the amount of warpage of the wafer is equal to or larger than the focal depth (about 1 μm) of the electron beam of the SEM.

If it takes a long time for the automatic focus adjustment processing, the throughput of the dimension measurement cannot be improved. Therefore, studies have been made to shorten the time required for the automatic focus adjustment processing by providing an optical distance measuring mechanism or the like. However, for the final focus adjustment, it is necessary to control an objective lens composed of an electromagnetic coil or the like, and a processing time of about 10 seconds is required for one focus adjustment. Therefore, if the automatic focus adjustment process is performed for each measurement point, it takes 10 seconds × the number of measurement points per wafer.

[0004]

As described above, as long as the automatic focus adjustment processing is performed for each measurement point, the throughput of the dimension measurement cannot be increased. In particular, when the diameter of a wafer increases in the future, the number of measurement points increases accordingly, and the measurement time per wafer increases.

Further, in the electron microscope, since the wafer is irradiated with an electron beam, the wafer surface is temporarily charged, and the floating positive ion impurities are attracted to the wafer surface to contaminate the wafer surface. Therefore, it is preferable that the irradiation amount of the electron beam is as small as possible. However, since the irradiation of the electron beam is also required for the automatic focus adjustment processing, the wafer surface is considerably contaminated when the automatic focus adjustment processing is performed for each measurement point.

The present invention has been made in view of the above points, and an object thereof is to improve the throughput by reducing the frequency of performing automatic focus adjustment processing as much as possible without lowering the precision of pattern recognition. Another object of the present invention is to provide an automatic focus control method and an automatic focus control system.

[0007]

According to a first aspect of the present invention, there is provided an automatic focusing apparatus capable of selecting whether or not to perform an automatic focusing process on a measurement object set on a microscope. In the control method, a power spectrum is obtained based on a result of performing a one-dimensional or two-dimensional Fourier transform process on the image of the measurement target captured by the microscope, and the power spectrum has a predetermined frequency or more. If the rate at which the frequency component is included is equal to or greater than a predetermined reference value, the pattern recognition processing of the measurement target is performed without performing the automatic focus adjustment processing. If the rate is less than the reference value, the pattern recognition processing is performed. Immediately before performing the above, the automatic focus adjustment processing is performed.

According to a second aspect of the present invention, there is provided an automatic focus adjusting means for performing automatic focus adjustment on a measurement object set on a microscope, and the measurement object based on an image of the measurement object taken into the microscope. In an automatic focus control system including pattern recognition means for performing pattern recognition processing,
Frequency component analysis means for obtaining a power spectrum based on a result of performing a one-dimensional or two-dimensional Fourier transform process on the image of the measurement object captured by the microscope, and a frequency higher than a predetermined frequency in the power spectrum. If the ratio of the components is equal to or more than a predetermined reference value, perform the pattern recognition processing by the pattern recognition unit without performing the automatic focus adjustment by the automatic focus adjustment unit, if the ratio is less than the reference value, An automatic focus control unit for performing automatic focus adjustment by the automatic focus adjustment unit immediately before performing the pattern recognition processing by the pattern recognition unit.

According to a third aspect of the present invention, in the automatic focus control system according to the second aspect, the automatic focus control means comprises:
The predetermined reference value is determined so that a detection error amount of a pattern obtained by the pattern recognition processing by the pattern recognition means is one pixel or less, which is a unit of image capture of the microscope.

According to a fourth aspect of the present invention, in the automatic focus control system according to the second or third aspect, the microscope is a scanning electron microscope, and the frequency component analyzing means is configured to detect different irradiation areas of the measurement object. Calculating the power spectrum, respectively, the automatic focus control means, for each of the irradiation area, determines whether or not the ratio of the power spectrum contains a frequency component of the predetermined frequency or more is not less than the reference value, For a portion where the ratio is equal to or more than the reference value, pattern recognition processing is performed by the pattern recognition unit without performing focus adjustment by the automatic focus adjustment unit. For a portion where the ratio is less than the reference value, the pattern recognition is performed. Immediately before performing pattern recognition processing by means, focus adjustment is performed by the automatic focus adjustment means.

According to a fifth aspect of the present invention, in the automatic focus control system according to any of the second to fourth aspects, the predetermined frequency is a frequency of 30 Hz or more.

According to a sixth aspect of the present invention, in the automatic focus control system according to any one of the second to fifth aspects, at least two arithmetic processing units are provided, one of which is dedicated to frequency analysis by the frequency component analysis means. Is provided.

The invention of claim 2 will be described in connection with, for example, the flowchart of FIG.
Corresponds to the processing in step S57, the "frequency component analysis means" corresponds to the processing in steps S52 and S53, and the "auto focus control means" corresponds to the processing in step S55.

[0014]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an automatic focus control system to which the present invention is applied will be specifically described with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of an automatic focus control system. The automatic focus control system shown in FIG. 1 uses a scanning electron microscope (SEM).
A length measuring system using croscopy is constituted, and is roughly divided into an electron beam column 1, a main body control device 2, and a host computer 3. Inside the sample chamber, a load lock chamber 11 for loading and unloading a sample (measurement target), an XY stage 12 on which the sample is mounted, an electron beam irradiation device 13 for irradiating the sample with an electron beam, A deflector 14 for controlling the irradiation direction and a secondary electron detector 15 for detecting secondary electrons from the sample are provided.

The main controller 2 includes an SEM controller 21 for controlling the deflector 14 and the secondary electron detector 15, and a stage controller 22 for controlling the moving direction and moving speed of the XY stage 12. The SEM control unit 21 outputs an analog signal (SEM image) corresponding to the amount of secondary electrons detected by the secondary electron detector 15.

The host computer 3 includes a main control unit 31 for controlling the entire system, an image processing unit 32 for analyzing an image of an object to be measured, and peripheral devices for controlling display of the display device 4 and controlling the XY stage 12. A control unit 33 and a storage device 34 are provided.

FIG. 2 is a block diagram showing a detailed configuration of the image processing unit 32 in the host computer 3 shown in FIG. The image processing unit 32 converts an SEM image output from the SEM control unit 21 into digital data by an A / D converter.
It includes a converter 51, a frame memory 52 for storing the converted digital data, an image processing operation unit 53 for performing a filtering process and the like, and a display control unit 54 for generating image data to be displayed on the display device 4 and the like. .

FIG. 3 is a main flowchart showing the processing operation of the automatic focus control system shown in FIG. 1. The operation of the present system will be described below with reference to this flowchart. First, the wafer to be measured is placed on the XY stage 12 in the sample chamber (Step S1). Next, the XY stage 12 is moved to perform the automatic focus adjustment processing at a reference point set in advance in the wafer (step S2), and the automatic focus adjustment processing is performed at this reference point (step S3).
In the best_focus state (focusing state), next, the image of the reference point is fetched and the pattern matching process is performed (step S4), and the alignment at the reference point (alignment adjustment) is performed. Details of this pattern matching processing will be described later.

Next, it is determined whether pattern matching processing has been performed for all the predetermined reference points, that is, whether or not the alignment adjustment has been completed (step S5). If there is any reference point that has not been adjusted yet, step S2 is performed. Return to

As described above, in steps S1 to S5, alignment adjustment is performed for each of a plurality of reference points preset in the wafer. Such adjustment is generally called global alignment adjustment.

When the pattern matching process is performed for all the reference points, the length measurement process after step S6 is performed.
In the length measurement process, first, the XY stage 12 is moved to a predetermined measurement point (step S6). Next, a process of determining whether or not to perform the automatic focus adjustment process at this measurement point (automatic focus execution availability determination process) is performed (step S7).

The purpose of the above-mentioned automatic focus adjustment processing is to enhance the accuracy and reliability of pattern recognition performed based on an SEM image from a sample, but the focus has already been adjusted to a level that does not affect the accuracy of pattern recognition. It is not uncommon that automatic focus adjustment processing takes time and is not very meaningful. Thus, in the automatic focus execution availability determination process, a process of determining in advance whether it is necessary to perform the automatic focus adjustment process is performed. The details of the automatic focus execution availability determination processing will be described later.

When the processing in step S7 is completed, next,
The image at the measurement point is captured and subjected to image analysis (pattern recognition processing) (step S8), and the pattern width is measured based on the result (step S9). Next, it is determined whether or not the length has been measured for all the measurement points (step S10). If there is a measurement point for which the length has not been measured yet, the process returns to step S6. Then, the wafer is unloaded (step S11).

As described above, steps S6 to S11 in FIG.
Then, the power spectrum is obtained based on the SEM image of each measurement point in the wafer, and the auto focus adjustment is omitted when the power spectrum includes a high frequency component equal to or more than a predetermined reference value. The frequency of processing can be reduced, and the throughput during measurement can be improved.

FIG. 4 is a detailed flowchart of the pattern matching process in step S4 of FIG. First,
The reference image data corresponding to the reference point is downloaded from the storage device 34 in the host computer 3 (Step S)
21). Next, an SEM image (hereinafter, referred to as measurement image data) from a reference point in the wafer is taken into the image processing unit 32 in the host computer 3 via the SEM control unit 21 (Step S22). The captured measurement image data is
This is analog data corresponding to secondary electrons from the wafer detected by the secondary electron detector 15.

Next, a spatial filtering process and a linear image enhancement process are performed (step S23). The spatial filtering process is a process for removing frequency components including noise, and the linear image enhancement process is a process for enhancing the contrast of a pattern.

Next, Fourier transform processing (FFT processing) is performed on the reference image data (template) downloaded in step S21 (step S24). Assuming that the function corresponding to the reference image data is g (m, n) and the function after the FFT processing is G (m, n), the relationship of equation (1) holds. Here, (m, n) indicates a pixel position as shown in FIG. 5, where 0 ≦ m ≦ 511 and 0 ≦ n ≦ 511.

[0029]

(Equation 1) Next, Fourier transform processing (FFT processing) is performed on the measurement image data captured in step S22 (step S25). The function corresponding to the measured image data is f (m,
n), and the function after FFT processing is F (m, n),
Equation (2) holds.

[0030]

(Equation 2) Next, a process (convolution process) of synthesizing the reference image data G (m, n) after the FFT process and the measurement image data (F (m, n)) is performed (step S26). m, n), the relationship of equation (3) holds.

[0031]

(Equation 3) Next, as shown in the equation (4), the function R (m, n) is calculated using the inverse F
The cross-correlation r (m, n) is obtained by FT processing (step S27).

[0032]

(Equation 4) The cross-correlation r (m, n) indicates the result of pattern matching, and is represented by a waveform as shown in FIG. Next, FIG.
, And the coordinates (m, n) at that time are detected,
The detection result is sent to the main controller 31 of the host computer 3 (Step S28).

As described above, in the pattern matching process, reference image data corresponding to a reference point is prepared in advance, and each of the measured image data and the reference image data is
The comparison is performed after T processing. Since the pattern matching process requires a complicated numerical operation, the image processing unit 32 instead of the main control unit 31 in the host computer 3 is used.
It is desirable to carry out. If the processing is performed by the image processing unit 32, the main control unit 31 can perform other processing even while the image processing unit 32 performs the calculation, and the processing efficiency of the entire system is improved. FIG. 7 is a detailed flowchart of the automatic focus execution availability determination processing in step S7 in FIG. First, the SEM corresponding to the measurement point
An image is captured (step S51). Imported SE
Since the M image is analog data, the A / D converter 51
Are stored in the frame memory 52.

Next, two-dimensional Fourier transform processing is performed on the image data stored in the frame memory 52 (step S52). Next, a frequency distribution (power spectrum) is obtained based on the two-dimensional Fourier transform processing result (step S53).

In an out-of-focus defocus state, the SEM image from the sample is apparently smoothed by the spread of the beam diameter, and the high-frequency component in the Fourier transform image obtained by performing the Fourier transform on the SEM image decreases. on the other hand,
In a Fourier transform image in a focused (best_focus) state, a signal component from a pattern mainly composed of low-frequency components and a noise component mainly composed of high-frequency components are mixed. Therefore, by analyzing the frequency component of the Fourier transform image, the defocus state and the best_focus state can be determined.

In step S53 of FIG. 7, a frequency distribution (power spectrum) is obtained from the square root of the sum of squares of the real part and the imaginary part of each frequency component of the Fourier transform image obtained by the two-dimensional Fourier transform processing. Then, the ratio of the high frequency component included in the power spectrum is detected (step S54).

FIG. 8A is a diagram showing a power spectrum in the best_focus state (or in the vicinity thereof), wherein the horizontal axis represents frequency [Hz] and the vertical axis represents intensity. The components below 30 Hz shown reflect the pattern shape of the wafer,
The 50-60 Hz component is due to noise from the power supply. The other high-frequency components are due to SEM-specific noise (also called sesame salt noise), and include, for example, noise associated with sampling.

FIG. 8B shows a power spectrum in a defocus state. In the defocus state,
As an effect of the spread of the beam diameter, an image obtained by apparently smoothing the SEM image is obtained, and the salt and pepper noise of the high frequency component shown in FIG. 8A is reduced. If pattern recognition processing is performed in the defocus state, pattern recognition may not be possible or incorrect pattern recognition may be performed.
It is desirable to perform automatic focus adjustment processing before performing pattern recognition.

Therefore, in step S55 of FIG.
When a high-frequency component other than the power supply frequency (50/60 Hz) is included at a predetermined reference value or more at 0 Hz or more,
It is determined that the state is the best_focus state, and the automatic focus adjustment processing is omitted. On the other hand, if the high frequency component of 30 Hz or more is less than the reference value, it is determined to be in the defocus state, and the automatic focus adjustment processing is performed (step S56).

Here, the reference value in step S55 depends on the performance of hardware such as the host computer 3 and the image processing unit 32 for performing the pattern recognition processing.
It is necessary to set the optimum value depending on the hardware used. As an example, when the image processing unit 32 employs an algorithm for detecting the maximum value of the normalized correlation coefficient, about 10%, preferably 10 to 15% of the total power spectrum intensity is set as the reference value. Is done.

When the processing in step S54 or S55 is completed, the pattern recognition processing in step S8 in FIG. 3 is performed.

In the embodiment described above, during the global alignment process (steps S1 to S5 in FIG. 3),
Although the automatic focus adjustment process is forcibly performed without performing the automatic focus execution availability determination process illustrated in FIG. 7, the automatic focus execution availability determination process may be performed during the global alignment process. As a result, the measurement time can be further reduced, and the throughput can be further improved.

In the above embodiment, an example using a scanning electron microscope (SEM) has been described.
It can be applied to other electron microscopes and optical microscopes.

As described above, the reference value used as a criterion for judging whether or not to perform the automatic focus adjustment processing varies depending on operating conditions such as hardware to be used. Changes and settings may be made. As a rough guideline of the reference value, it is desirable that the result of the pattern recognition be set so that the error of the reference pattern is equal to or smaller than a subpixel.

Further, it may be possible to arbitrarily select whether or not to perform the automatic focus execution availability determination processing shown in FIG. Further, instead of performing the automatic focus adjustment processing, a mode for manually performing the focus adjustment may be provided.

In the above-described embodiment, an example in which the SEM image is subjected to the two-dimensional Fourier transform processing has been described. However, the power spectrum may be obtained based on the result of the one-dimensional Fourier transform processing.

In the above-described embodiment, an example in which the dimension of a fine pattern in a wafer is measured has been described. Can also be applied.

[0048]

As described in detail above, according to the present invention, if the rate at which the frequency component higher than the predetermined frequency is included in the power spectrum corresponding to the image captured by the microscope is smaller than the predetermined reference value. Since the pattern recognition processing of the measurement target is performed without performing the automatic focus adjustment processing,
The measurement time of the measurement target can be reduced, and the throughput can be improved. In particular, the present invention exerts a remarkable effect when the number of measurement points per measurement object is large, such as a large diameter wafer.

When an electron microscope is used as a microscope, it is necessary to irradiate an electron beam to the object to be measured, so that impurities easily adhere to the surface of the object to be measured. Since the number of times of performing the focus adjustment processing can be reduced, the irradiation amount of the electron beam can be reduced, and the contamination of the surface of the measurement target can be reduced.

[Brief description of the drawings]

FIG. 1 is a block diagram of an embodiment of an automatic focus control system.

FIG. 2 is a block diagram showing a detailed configuration of an image processing unit in the host computer shown in FIG. 1;

FIG. 3 is a main flowchart showing a processing operation of the automatic focus control system.

FIG. 4 is a detailed flowchart of a pattern matching process in step S4 of FIG. 3;

FIG. 5 is a view for explaining coordinates of reference image data.

FIG. 6 is a waveform diagram of a cross-correlation r (m, n).

FIG. 7 is a detailed flowchart of an automatic focus execution availability determination process in step S7 of FIG. 3;

8A is a power spectrum diagram in a best_focus state, and FIG. 8B is a power spectrum diagram in a defocus state.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 electron beam column 2 main body control device 3 host computer 4 display device 11 load lock room 12 XY stage 13 electron beam irradiation device 14 deflector 15 secondary electron detector 21 SEM control unit 22 stage control unit 31 main control unit 32 image Processing unit 33 Peripheral control unit 34 Storage device

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G02B ^7/ 28-7/40 G03B 13/36

Claims (6)

(57) [Claims] 1. An automatic focus control method capable of selecting whether or not to perform an automatic focus adjustment process on a measurement target set on a microscope, wherein the automatic focus control method includes: A power spectrum is obtained based on the result of the one-dimensional or two-dimensional Fourier transform processing. If the ratio of the power spectrum containing a frequency component equal to or higher than a predetermined frequency is equal to or higher than a predetermined reference value, the automatic focus adjustment is performed. Performing the pattern recognition processing of the measurement target without performing the processing, and if the ratio is less than the reference value, performing the automatic focus adjustment processing immediately before performing the pattern recognition processing. . 2. An automatic focus adjusting means for performing automatic focus adjustment on a measurement target set on a microscope, and performing a pattern recognition process on the measurement target based on an image of the measurement target taken into the microscope. An automatic focus control system including a pattern recognition unit, a frequency component analysis for obtaining a power spectrum based on a result of performing a one-dimensional or two-dimensional Fourier transform process on the image of the measurement target captured by the microscope. Means, if the rate at which the power spectrum contains a frequency component equal to or higher than a predetermined frequency is equal to or higher than a predetermined reference value, perform pattern recognition processing by the pattern recognition unit without performing automatic focus adjustment by the automatic focus adjustment unit. If the ratio is less than the reference value, immediately before performing the pattern recognition process by the pattern recognition means. Automatic focus control system characterized by comprising an automatic focus control means for performing automatic focus adjustment by serial automatic focusing means. 3. The automatic focus control means according to claim 1, wherein said predetermined reference value is set such that a detection error amount of a pattern obtained by pattern recognition processing by said pattern recognition means is one pixel or less, which is an image capturing unit of said microscope. The automatic focus control system according to claim 2, wherein: 4. The microscope according to claim 1, wherein the microscope is a scanning electron microscope, wherein the frequency component analyzing means obtains the power spectrum for each of a plurality of different positions of the object to be measured, and wherein the automatic focus control means comprises: In each case, it is determined whether or not the ratio of the frequency spectrum containing the frequency component equal to or higher than the predetermined frequency is equal to or higher than the reference value. A pattern recognition process is performed by the pattern recognition unit without performing the adjustment, and for a portion where the ratio is less than the reference value, focus adjustment is performed by the automatic focus adjustment unit immediately before performing the pattern recognition process by the pattern recognition unit. The automatic focus control system according to claim 2 or 3, wherein: 5. The automatic focus control system according to claim 2, wherein said predetermined frequency is a frequency of 30 Hz or more. 6. The automatic focus control according to claim 2, further comprising at least two arithmetic processing units, one of which is provided exclusively for frequency analysis by said frequency component analysis means. system.